Skip to main content

Table 4 External evaluation results on the IMPRESS [24] dataset (\({\text{val}}_{\rm ext}\)): AUCs, p values and confidence intervals for predicting pCR

From: PROACTING: predicting pathological complete response to neoadjuvant chemotherapy in breast cancer from routine diagnostic histopathology biopsies with deep learning

Biomarker

AUC

p

95% CI

TNBC N = 64 (Npcr = 27)

 IMPRESS (H&E only)

0.698

–

–

 cTILs

0.708

0.005

0.574–0.841

 LTR

0.686

0.012

0.550–0.821

 ITR

0.715

0.004

0.584–0.845

 MTR

0.608

0.144

0.468–0.747

HER2+ N = 62 (Npcr = 38)

 IMPRESS (H&E only)

0.812

–

–

 cTILs

0.679

0.018

0.543–0.814

 LTR

0.682

0.016

0.544–0.820

 ITR

0.664

0.016

0.527–0.802

 MTR

0.615

0.129

0.475–0.756